Sei Giga Roadmap Targets 200K TPS and Sub-400ms Finality

The roadmap site went live on May 27, 2026, at giga.seilabs.io. One day later, Sei’s official X account posted “Introducing The Giga Roadmap.” The post included a milestone tracker and video explainer. The announcement triggered a roughly 15% price surge in $SEI within 24 hours, according to CoinMarketCap data.

Sei Giga is not a single upgrade. It is a multi-phase engineering program that rebuilds Sei’s execution client, storage layer, and consensus mechanism from scratch. Full EVM compatibility is preserved throughout.

What Sei Giga Actually Targets

The program sets three core performance goals. First, 5 gigagas per second of throughput. Sei Labs claims this is roughly 50 times faster than any current EVM mainnet. Second, more than 200,000 TPS for certain workloads. Third, sub-400 millisecond finality.

Sei’s current mainnet already delivers roughly 400 ms block times through its Twin Turbo consensus. The Giga upgrades aim to push finality even lower through a redesigned consensus layer.

These numbers come from an internal devnet that Sei Labs stress-tested, according to Sei’s official documentation. No public mainnet has achieved these figures yet. The targets remain benchmarks rather than live performance metrics.

Sei Giga Builds on the v2 Foundation

Sei launched its v2 upgrade in July 2024. That release introduced parallelized EVM execution through Optimistic Concurrency Control (OCC). This method processes transactions simultaneously instead of one at a time. While v2 improved throughput, it served as the foundation for a larger overhaul.

In mid-December 2024, Sei Labs formally announced the Giga vision. A blog post on December 18 outlined three workstreams: Execution, Consensus, and Storage. A press release followed on December 20. Around May 2025, the team published the Giga Whitepaper v1 to formalize the architecture.

Key design choices include asynchronous execution, which decouples consensus from state changes. The storage overhaul swaps Merkle trees for a flat key-value store with lattice hashes. And the new multi-proposer BFT consensus allows every validator to propose blocks in parallel “lanes.”

Current Roadmap Progress

According to the public roadmap, several milestones are already complete. The Giga Whitepaper v1 defined the core specifications. In addition, the internal devnet demonstrated 5 gigagas and 200,000+ TPS under controlled conditions.

Three major upgrades are actively in progress as of late May 2026. The Ares upgrade rebuilds the execution client from scratch. It adds pipeline parsing, validation offloading, and VM runtime optimizations. Meanwhile, the Eidos upgrade overhauls storage. It replaces Merkle trees with a flat KV store and adds hybrid data tiers.

SIP-3, the EVM-only transition proposal, represents a decisive pivot. It moves all staking to EVM, disables IBC transfers, and deprecates CosmWasm. Mainnet v6.5 upgrades went live in May 2026. As a result, exchanges and custodians faced a June 1, 2026, deadline for migration support.

Upgrades Still on the Horizon

The roadmap lists several upgrades under “Coming Soon” without firm dates. Autobahn introduces multi-proposer BFT consensus. In this design, every validator proposes transactions in parallel lanes. It also delivers order-first processing and asynchronous state commitments to reach 200,000+ TPS.

Sedna adds a private mempool designed to combat MEV extraction. In this system, transaction fragments spread across lanes and reassemble after ordering.

Hermes represents the next-generation consensus layer, though a whitepaper for it is still forthcoming. Sei Labs also plans a Whitepaper v2 with further performance refinements.

SIP-3 Forces a Strategic Choice

SIP-3 carries significant implications for the existing Sei ecosystem. By disabling IBC, Sei cuts off its Cosmos interoperability roots. Consequently, Cosmos-native assets like USDC.n must migrate before the governance vote or risk becoming inaccessible.

This is a deliberate tradeoff. Sei is going all-in on EVM to simplify its stack. Developers using Solidity, JSON-RPC, and standard EVM tooling face zero migration friction. But projects built on CosmWasm or relying on IBC must adapt or leave.

Sei co-founders have called the Sei Giga program the hardest engineering challenge since Ethereum’s Merge. The rollout follows an incremental approach through 2026, with testnet validation before each mainnet deployment.

SEI Price Reacts to Roadmap News

$SEI traded in the $0.06 to $0.07 range through late May 2026, according to CoinMarketCap data. On May 28, whale accumulation and the roadmap announcement drove 24-hour trading volume to approximately $133 million.

Despite the short-term bounce, $SEI remains down between 38% and 50% year-to-date, depending on the entry point. The broader altcoin market and delays in full Giga delivery have weighed on sentiment throughout 2026.

Analyst forecasts for $SEI by year-end 2026 range from $0.05 to $0.40, according to aggregated estimates on TradingView. Those projections depend heavily on Sei Giga milestones and market conditions. This is not financial advice.

What Remains Unknown

Sei Giga has clearly defined technical specifications and a transparent milestone tracker. Yet several critical unknowns remain.

The roadmap uses “Coming Soon” labels instead of hard dates for Autobahn, Ares, Eidos, Sedna, and Hermes. Real-world mainnet performance after Autobahn may differ from devnet benchmarks. Live networks carry economic activity, validator diversity, and unpredictable transaction patterns.

Adoption impact also remains uncertain. TVL growth, dApp migration rates, and institutional inflows will depend on execution quality and market conditions. Final MEV mitigation strategies and any specification changes in Whitepaper v2 could also shift the roadmap direction.

A High-Stakes Bet on Monolithic Speed

Sei Giga represents one of the most ambitious technical bets in the current L1 landscape. Ethereum scales through Layer-2 rollups and data sharding. Sei takes the opposite approach, rebuilding a single base layer for raw throughput.

The competition is fierce. Monad, MegaETH, and Pharos are all pursuing high-performance EVM execution. The 5-gigagas claim is the most aggressive target on record. But delivering it on a live mainnet with real users differs fundamentally from a controlled devnet demonstration.

Whether the engineering delivers will determine if Sei claims the performance crown it is chasing. The upgrades roll out incrementally. The next major test comes when Autobahn reaches testnet.